The antigen binding domain of an antibody comprises two separate regions: a heavy chain variable domain (VH) and a light chain variable domain (VL: which can be either Vkappa or Vlambda). The antigen binding site itself is formed by six polypeptide loops: three from VH domain (H1, H2 and H3) and three from VL domain (L1, L2 and L3). A diverse primary repertoire of V genes that encode the VH and VL domains is produced by the combinatorial rearrangement of gene segments. The VH gene is produced by the recombination of three gene segments, VH, D and JH. In humans, there are approximately 51 functional VH segments (Cook and Tomlinson (1995) Immunol Today, 16: 237), 25 functional D segments (Corbett et al. (1997) J. Mol. Biol., 268: 69) and 6 functional JH segments (Ravetch et al. (1981) Cell, 27: 583), depending on the haplotype. The VH segment encodes the region of the polypeptide chain which forms the first and second antigen binding loops of the VH domain (H1 and H2), whilst the VH, D and JH segments combine to form the third antigen binding loop of the VH domain (H3). The VL gene is produced by the recombination of only two gene segments, VL and JL. In humans, there are approximately 40 functional Vk segments (Schäble and Zachau (1993) Biol. Chem. Hoppe-Seyler, 374: 1001), 31 functional Vλ segments (Williams et al. (1996) J. Mol. Biol., 264: 220; Kawasaki et al. (1997) Genome Res., 7: 250), 5 functional Jκ segments (Hieter et al. (1982) J. Biol. Chem., 257: 1516) and 4 functional Jλ segments (Vasicek and Leder (1990) J. Exp. Med., 172: 609), depending on the haplotype. The VL segment encodes the region of the polypeptide, chain which forms the first and second antigen binding loops of the VL domain (L1 and L2), whilst the VL and JL segments combine to form the third antigen binding loop of the VL domain (L3). Antibodies selected from this primary repertoire are believed to be sufficiently diverse to bind almost all antigens with at least moderate affinity. High affinity antibodies are produced by “affinity maturation” of the rearranged genes, in which point mutations are generated and selected by the immune system on the basis of improved binding.
The heavy chain locus contains a large number of variable chain genes (VH; in fact not complete genes but comprising a first coding exon plus transcriptional start site) that are recombined onto two short coding regions D and J (known as VDJ recombination) which procede the exons that code for the constant region of the heavy chain Cμ to give a complete antibody heavy chain gene known as IgM. Subsequently a class switch takes place where the variable part is recombined with another constant region that is located downstream of the IgM constant region to give IgD, IgG, IgA and IgE (coded for by the exons of the various Cδ,Cγ,Cα,Cε located downstream of the gene for Cμ. The intervening constant regions are deleted in the process. A similar process takes place in the light gene loci, first the κ locus, and when this does not lead to a productive antibody in the λ locus (for review see Rajewski, K., Nature 381, p751-758, 1996; for an extensive review, see the textbook Immunobiology, Janeway, C., Travers, P., Walport, M., Capra. J., Current Biology Publications/Churchill Livingstone/Garland Publishing, fourth edition, 1999, ISBN 0-8153-3217-3).
Camelids (camels, dromedary and llamas) contain, in addition to normal heavy and light chain antibodies (2 light chains and 2 heavy chains in one antibody), single chain antibodies (containing only heavy chains). These are coded for by a distinct set of VH segments referred to as VHH genes. Antigen binding for single chain antibodies is different from that seen with conventional antibodies, but high affinity is achieved the same way, i.e. through hypermutation of the variable region and selection of the cells expressing such high affinity antibodies (affinity maturation). The VH and VHH are interspersed in the genome (i.e. they appear mixed in between each other). The identification of an identical D segment in a VH and VHH cDNA suggests the common use of the D segment for VH and VHH. Natural VHH containing antibodies are missing the entire CH1 domain of the constant region of the heavy chain. The exon coding for the CH1 domain is present in the genome but is spliced out due to the loss of a functional splice acceptor sequence at the 5′side of the CH1 exon. As a result the VDJ region is spliced onto the CH2 exon. When a VHH is recombined onto such constant regions (CH2, CH3) an antibody is produced that acts as a single chain antibody (i.e. an antibody of two heavy chains without a light chain interaction). Binding of an antigen is different from that seen with a conventional antibody, but high affinity is achieved the same way, i.e. through hypermutation of the variable region and selection of the cells expressing such high affinity antibodies.
The structure of isolated VH domains has been determined using NMR and X-ray crystallography techniques (Spinell et al, (1996), Nat Structural biol. 3, 752). Data show that the Immunoglobulin fold is well preserved in Camelid VHH domains. Two beta sheets (one with four and one with five beta-strands) are packed against each other and stabilised by a conserved intradomain disulphide bond between C22 and C92. The side of the camel VHH domain corresponding to the VL interface of the normal VH in an Fv has a quite different architecture. Compared to the human VH, four amino acid substitutions are located in this region.
From a survey of all human and mouse VH antigen binding loop structures, it is apparent that there are only a restricted number of possible conformations. Three and four different conformations are described for the first and second antigen binding loop respectively. These canonical structures are determined by the length of the loop and the presence of particular residues at key positions. The H3 loop is extremely variable in length and sequence (Wu et al (1993) Proteins: structure, funct and genet., 16, 1). Surprisingly, the antigen binding loop of camel VH domains deviate from the canonical loop definitions of human and mouse VHH domains. This deviation could not be predicted as the loop length and the residues at the key positions are very similar between camel VH and human VH. The additional canonical loop structures in camel VH domains make the structural repertoire of their paratope larger than that of VH domains in Fv fragments from conventional antibodies. Moreover, the hypervariable region around the first antigen binding loop is enlarged compared with human or mouse antibodies. It is thought that the extension of the first hypervariable region and concomitant enlarged antigen binding surface compared to that of a VH in a conventional antibody compensates in part for the absence of a VL domain (Riechmann, L. & Muyldermans, S (1999), 231 25-38).
A single domain camelid VHH antibody as well as being more suitable for structural analysis than the larger heavy and light chain antibody molecules, also provides a small and efficient antigen binding unit. Such an antibody has many and varied therapeutic potential. In addition, it has been found that camelid single chain antibodies can bind antigens which are inaccessible to antibodies possessing both heavy and light chains. It is thought that this ability is due to the presence of a large protruding third hypervariable loop of 10 amino acids or more which can insert into cavities of antigen surfaces. This is especially significant as the catalytic site of an enzyme is often located at the largest cavity on their protein surface. (Ladowski, R. A (1996). Protein Science 5, 2438). Such sites are not normally immunogenic for conventional antibodies (Novotny, J et al, (1986) Proc Nat Acad Sci USA, 83, 226). In the structure of the camel VHH cAb-Lys3, the 24 residue H3 loop penetrates deeply into the active site of lysozyme (Transue, T. R et al (1998) Prot: Structure, Funct and Genet, 32, 515), showing that Camel heavy chain antibodies have the potential to form specific enzyme inhibitors.
Recently, isolated Camelid VHH domains have been generated in bacteria (Riechmann, L et al. Journal of Immunological Methods 231 (1999), 25-38). However bacterial expression systems have the disadvantage that they do not perform post-translational modifications. Such modifications, in particular glycosylation events, are crucial for the effective functioning of antibodies, particularly in an in vivo environment.
In the same study, the genes for Camel VHH domains were inserted into expression vectors and expressed in Cos cells to generate multi-domain proteins. In one example, an intact single heavy chain only antibody was generated by cloning a particular camel VHH in front of the hinge and effector function domains of human IgG1. The expression in Cos cells has the advantage over bacterial expression systems that post-translational modification events occur in these cells. Consistent with this was the finding that these antibodies were fully active in antigen binding. The DNA for the generation of these constructs is generally isolated from mature (ie those which have undergone affinity maturation) antibodies generated from B cells. Although these single chain antibodies expressed in mammalian cells in an in vitro environment can bind to one or more antigens, they cannot undergo the processes of class (isotype) switching and affinity maturation (hypermutation). Thus the single chain antibodies expressed in Cos cells do not undergo the process of antibody evolution as those naturally occurring antibodies generated within a mammal. It is this process of antibody evolution which results in the production of specific antibodies which bind with high affinity. Thus, there remains a need in the art for a method allowing the generation of single chain VHH antibodies, in a mammal such that the normal processes of antibody evolution can take place.
In addition Camelid single chain antibodies have also been selected and expressed using phage display technology. (Riechmann, L. & Davies, S. J. Biomol. NMR, 6, 141). Again though, the antibody constructs are generated from nucleic acid isolated from mature B cells or spleen, and therefore as with the case above, the antibodies expressed do not undergo class switching and somatic hypermutation (affinity maturation) which is necessary for the production of specific antibodies which bind to their antigen with selectivity and high affinity.
The present inventors realised that if they could understand the mechanism by which camelid single chain antibody molecules evolve (by class-switching and affinity maturation) during early antibody development in B cells, then this system may be recreated in vivo. This would allow the generation of vast quantities of an evolved single chain antibody for structural, therapeutic and diagnostic applications.